Xiaoping P. Liu

Similarity measures for content-based image retrieval based on intuitionistic fuzzy set theory

In this paper, a new intuitionistic fuzzy model for images based on the HSV color histogram is proposed. The image can be treated as an Attanassov’s intuitionistic fuzzy set (IFS) with this new model. A new and simple calculation of similarity measurement called IFSL1 based on similarity measurement of intuitionistic fuzzy set L1 is presented. Unlike general fuzzy similarity measure that consider only the membership degree, the new intuitionistic similarity measure takes into account the membership degree, the nonmembership degree and the hesitation degree, these have been found to be highly useful in dealing with vagueness. The similarity measure IFSL1 is used for content-based image retrieval (CBIR).With the similarity measure IFSL1, image retrieval can be carried out more rapidly than with many other existing similarity measurements and the results better coincide with human perception.

An improved realistic mass-spring model for surgery simulation

An improved realistic mass-spring model, which is mainly based on the 3D finite strain nonlinear anisotropic elasticity theory, is presented for virtual reality based surgery simulation. Compared with the conversional mass-spring model, the proposed model is able to describe typical behaviors of living tissues such as incompressibility, nonlinearity and anisotropy. The nonlinear viscoelasticity is also incorporated into the soft tissue model by employing a numerical scheme. In terms of implementation, the model proposed can be seen as a mixture of finite-element and mass-spring models, which enables it to still maintain the advantage of mass-spring model, such as simple architectures, low memory usage and fast computation. An example to use this model to simulate human kidney is given to demonstrate its capability of describing the typical behaviors of soft tissue.

Stochastic Subset Selection for Learning With Kernel Machines

Kernel machines have gained much popularity in applications of machine learning. Support vector machines (SVMs) are a subset of kernel machines and generalize well for classification, regression, and anomaly detection tasks. The training procedure for traditional SVMs involves solving a quadratic programming (QP) problem. The QP problem scales super linearly in computational effort with the number of training samples and is often used for the offline batch processing of data. Kernel machines operate by retaining a subset of observed data during training. The data vectors contained within this subset are referred to as support vectors (SVs). The work presented in this paper introduces a subset selection method for the use of kernel machines in online, changing environments. Our algorithm works by using a stochastic indexing technique when selecting a subset of SVs when computing the kernel expansion. The work described here is novel because it separates the selection of kernel basis functions from the training algorithm used. The subset selection algorithm presented here can be used in conjunction with any online training technique. It is important for online kernel machines to be computationally efficient due to the real-time requirements of online environments. Our algorithm is an important contribution because it scales linearly with the number of training samples and is compatible with current training techniques. Our algorithm outperforms standard techniques in terms of computational efficiency and provides increased recognition accuracy in our experiments. We provide results from experiments using both simulated and real-world data sets to verify our algorithm.

An improved switching vector median filter for image-based haptic texture generation

In this paper, we present a novel approach to generation of haptic texture from visual image for modelling the constraint forces in tangent direction of a surface, based on which the texture force (friction) can be calculated. One significant difference from conventional image-based haptic texture is that an improved switching vector median filter (ISVMF) was employed to replace the gray-scale Gaussian filter for preserving fine detail information of the image. As a result, the new haptic texture rendering algorithm can convey tactile patterns based on fine features of the image and is much more realistic than those published in the literature.

Adaptive kernels for data recovery in tele-haptic and tele-operation environments

The development of non-linear filtering through the use of kernel machines has gained much popularity in recent years. Both the kernel least-mean-squared (KLMS) [1], and kernel recursive least-mean-squared (KRLS) [2] have been used to provide superior regression performance to traditional linear methods. As well, there has been developments in on-line support vector machine techniques that allow non-linear regression to previously off-line batch methods [3] [4]. In this paper we present a novel adaptive method for tuning the kernel parameter of a Gaussian kernel when using the KLMS algorithm. We test our algorithm on both simulated, and real data captured from a haptic device.

Human-robot interaction via haptic device

This paper presents a novel human-robot interaction system for which haptic interaction is involved. The system consists of a nonholonomic mobile robot and a Phantom Omni haptic device. The tracking problem on the 2D plane is analyzed and the backstepping technique is used for designing the tracking algorithm for the nonholonomic mobile robot. A haptic rendering algorithm is designed to generate the haptic feedback based on the position error information. Experiments show that the user can guide the movement of the robot quite easily and smoothly.

The partitioned kernel machine algorithm for online learning

Kernel machines have been successfully applied to many engineering problems requiring pattern recognition and regression. Kernel machines are a family of machine learning algorithms including support vector machines (SVM) [1], kernel least mean squares adaptive filter (KLMS) [2], and kernel recursive least squares (KRLS) adaptive filter [3] to name a few. In this paper we present the partitioned kernel machine algorithm for use in online learning in virtual environments. The PKM algorithm enhances the accuracy of the computationally efficient KLMS algorithm. The PKM algorithm is an iterative update procedure that focuses on a subset of the stored vectors in the kernel machine buffer. We use a similarity measure for the selection of kernel machine vectors that allow more common vectors to be updated more frequently, and outlier vectors to be updated less frequently. We validate the increased accuracy of our novel algorithm in two separate experimental settings.

Stability analysis of haptic shared control and its application on a mobile assistive robot

Haptic shared control (HSC) has been widely used in human-robot interaction systems, such as car driving and assistive robots. However, the stability analysis of HSC is not well investigated in current literature, especially for the nonlinear systems. In this paper, the stability of HSC for nonlinear systems is analyzed based on the set theory and convex analysis. The authority weight calculation and haptic rendering algorithm of HSC are developed based on the stability analysis. Then, we apply these algorithms to our assistive robot system. Experimental results show the stable operation of the assistive robot and the effectiveness of the haptic rendering algorithm.

Truncation Error Compensation in Kernel Machines

The analysis and prediction of time series data has played an important role for intelligent systems used in the area of cybernetics and human-machine interaction. Time series prediction is especially important in the case of unreliable communication of data acquired by intelligent systems. Computationally efficient kernel based regression algorithms have allowed for the prediction of non-linear relationships within time series data. In this paper, we present the smooth delta corrected kernel least mean square (SDC-KLMS) algorithm. The SDC-KLMS scales in linear time with the number of samples stored, hence making it computationally efficient. We present a theoretical motivation for our algorithm and we experimentally show how our approach overcomes a limitation imposed by the use of a finite storage buffer. Experiments with simulated, benchmark, and real world data were conducted to verify the accuracy of our algorithm.